Interview with Raj Karnani, Healthcare Analytics Translator, Karnani Healthcare Analytics

Healthcare has become more important today than ever before. The industry has been innovating and growing digitally at a rapid pace to advance various areas of telemedicine, patient monitoring & engagement, drug discovery, and many others to cope up with the population health demand.

In all the above, data analytics plays a very key part in understanding the business insights and prepare for the future, and who better to tell us than a pioneer in the field of healthcare analytics.

We had the pleasure of interviewing Raj Karnani from Karnani Healthcare Analytics and we thank him for taking part in the Data Science Interview Series and sharing several insights, including:

  • His journey from a physician to a clinical data scientist
  • Growth of digitalization in the healthcare industry
  • Key developments in healthcare analytics
  • Bridging the gap between business and IT
  • Advice to data science aspirants

#Getting into Data Science

While learning math and coding is certainly important, don’t underestimate the importance of communication.

Raj Karnani

How did you first get into data science? When did the journey begin?

As a practicing physician, I noticed that most diseases present with the same patterns over and over again. What separates an expert physician from a novice is the ability to recognize these patterns quickly and accurately. When electronic medical records (EMRs) appeared, I wondered if this tool could help enhance physicians’ ability to recognize disease patterns. EMRs, unfortunately, was designed primarily for patient billing, and they had significant limitations in performing analytics. It is at that point I wondered if EMRs could be re-engineered. After more research, I eventually made the decision to pursue coursework in data analytics and become a clinical data scientist.

What are the key skills that you use every day as a data scientist, and how did you develop them?

There are many skills I use daily as a clinical data scientist. These include critical thinking, teaching, project management, product management, and communication. I developed many of these skills while practicing as an academic physician where my focus was on teaching medical students and resident physicians, caring for and communicating with patients, and conducting quality improvement projects.

Then, with some recent analytics coursework and practice doing analytics projects, I was able to round out my skill set. While learning math and coding is certainly important, don’t underestimate the importance of communication. Data science projects are always done in teams, and team members must be on the same page. Moreover, the consumers of your data projects will rarely be familiar with the intricacies of data science, so you have to continually practice ways to effectively communicate on their level.

What are the top challenges you currently face as a professional data scientist, and how do you go about tackling them?

There are many challenges I face, but the most important one is to turn the ever-increasing amounts of data we collect in healthcare into information that is usable and actionable. In order to exist as a profession, we have to show value to our stakeholders. In other words, by analyzing healthcare data with sophisticated machine learning algorithms, we have to show how we are improving population health and reducing costs. Each stakeholder has unique priorities, and these priorities continually change, so being nimble and working to create and show value is always challenging.


#Data Science at the workplace

Domain knowledge of the industry you work in is paramount to becoming a highly successful data scientist, and it is what separates the best from the rest.

Raj Karnani

How important is the domain knowledge of the business/industry you’re in as a data scientist, and how did you acquire it?

Domain knowledge of the industry you work in is paramount to becoming a highly successful data scientist, and it is what separates the best from the rest. I acquired knowledge of the medical industry from practicing as a physician prior to transitioning to data science in healthcare. This is not possible for everyone, though. Many data scientists enter the field straight out of college after majoring in math or engineering. In these instances, I recommend finding people either at your place of work or in your professional network who are experts in your industry that you can turn to for questions you have about the industry.

In other words, have a reliable way to get the industry knowledge you need. If you focus exclusively on the math as you start your career, it will be difficult to rise up the data science ranks.

Data Science is penetrating every industry today. How has it affected the growth of the healthcare industry?

Data science is the next frontier in medicine. It hasn’t penetrated the world of medicine and healthcare like it has other industries, but it will over the next 5-10 years, and it will likely lead to substantial growth. Medicine, as an industry, is typically about 15-20 years behind other industries when it comes to adopting new technologies. Part of this is due to the general reluctance healthcare workers have in using data to guide decision-making.

Heavily siloed data, lack of interoperability of various electronic medical records between and even within the same hospital, and patient privacy concerns have added to this difficulty. While most of the existing data science in healthcare has focused on structured data like labs and vital signs, the vast majority of information is contained within unstructured data, namely, written notes.

Natural language processing will be a big area of exploration in trying to extract this largely untapped data.

What is the biggest data science component that you have incorporated in your workflow in the last 12 months that has created an impact?

The biggest data science component for me isn’t really a component, it’s the ability to take a step back and think through the business problem carefully so that we are coming up with a solution that will work and will make the lives of physicians and nurses better. New tools and programs are great, but they will always change. If you chase tools, you will always be chasing the next popular thing. Instead, chase the skill of problem solving. Often times, you don’t need a sophisticated data science component to solve a problem. Many problems can be solved with simple algorithms and/or diagnostic analytics.


#Business, IT and Bridging the gap

Find your passion and what you want to change, then become a data scientist.

Raj Karnani

Do you see a big gap between the business and IT? What steps would you take to bridge the gap?

In healthcare, there is definitely a gap between medical practitioners and IT, and it is this gap I try to bridge as an analytics translator. One of the big reasons that physicians are generally disgruntled with electronic medical records is that they were not heavily involved in their design and creation.

As a result, electronic medical records are primarily helpful only for providing documentation for billing at the expense of worsening workflow. Physicians need to be heavily involved in all applications of IT, including data science and artificial intelligence. The best way to do this is to have cadre of physicians who are cross-trained in machine learning and AI so they can bridge the gap between the business (i.e., physicians) and IT.

What advice would you give to someone who wants to get into data science today? What roadmap should one follow to become a data translator?

My advice to someone who wants to get into data science today is to develop domain knowledge first. In other words, find your passion and what you want to change, then become a data scientist. Data science is the convergence of three areas: math, computer science, and domain knowledge. Most data scientists are good in math and computer science, but not with domain knowledge. It is very difficult to be good in all three areas, but if you start with domain knowledge, you will be able to put yourself head and shoulders above all others as you gain knowledge of math and computer science.

In college, I recommend pursuing a double major – one in a field that is geared toward data science (i.e., math, computer science, engineering), and another in a domain of interest (finance, manufacturing, healthcare, etc.). That way, you can develop data translator skills naturally as you begin your career.

About The Author

Scroll to Top
Share via
Copy link
Powered by Social Snap